Skip to content
Smeet Patel.

AI & GenAI Engineering · Available for projects

Production AI,
not prototypes.

I'm Smeet — an AI & GenAI Engineer and Technical Lead with 7+ years of software engineering experience. I build AI agents, RAG systems and LLM applications that connect to real business processes, APIs and SaaS platforms.

Hire me on UpworkWhatsApp
Years engineering
7+Years engineering
Engineers led
15Engineers led
Apps delivered
12+Apps delivered
Enterprise users
22,000+Enterprise users

Ahmedabad, India · Working with clients worldwide

What I actually do

My focus is turning AI from a standalone feature into a reliable workflow — automating document processing, connecting AI to business systems, building agentic workflows, and integrating LLMs into applications that already exist.

If you have an AI workflow that needs to move from idea or prototype to a working production system, I can help design the architecture and build the implementation.

Sounds familiar?

Three ways AI projects stall before they ship.

  • 01

    The prototype works. It just won't productionise.

    A demo that impressed everyone now has to handle real traffic, real edge cases and real data — and nobody is sure what it takes to get there.

  • 02

    AI was bolted on, and it never reached your systems.

    There's a model in the product, but it doesn't touch the CRM, the database or the documents where the actual work happens.

  • 03

    Your team still processes documents by hand.

    Invoices, contracts, forms and reports move through people who could be doing higher-value work, because nothing automated them reliably enough to trust.

What I build

Types of projects I have delivered.

Nine categories of work, from AI agents and RAG systems through to the cloud infrastructure that keeps them running.

  • AI & GenAI Applications

    AI agents, RAG systems and LLM applications built to run in production — grounded in your data, wired to your APIs, and observable when they fail.

    AI AgentsRAGOpenAI APILangChain

  • AI-Powered SaaS

    Scalable SaaS platforms with AI capabilities, subscription workflows and the deployment pipeline to ship them continuously.

    SaaSSubscriptionsMulti-tenant

  • Workflow Automation

    Document-heavy and manual business processes turned into automated pipelines that connect LLMs to the systems your team already uses.

    Document IntelligencePipelinesIntegrations

  • Full-Stack Web Applications

    End-to-end product development across frontend, backend, database and infrastructure — architecture through production support.

    ReactNext.jsNode.jsNestJSPython

  • Enterprise Applications

    Production applications inside enterprise environments, built to strict security, compliance, code-review and CI/CD standards.

    Fortune 50ComplianceAccessibility

  • Real-Time Applications

    WebSocket-based communication and real-time systems that hold up under concurrent connections with low latency.

    WebSocketsSocket.IOLow latency

  • Subscription & Payment Systems

    Subscription billing, automated payment workflows and referral mechanics — the parts that must never quietly break.

    BillingPayment gatewaysReferrals

  • AWS Cloud Solutions

    EC2, S3, CloudFront, Lambda and RDS for hosting, asset delivery, serverless processing and relational persistence.

    EC2S3CloudFrontLambdaRDS

  • Blockchain & Web3

    Blockchain integrations, Solidity smart contracts and dApps connected into conventional application stacks.

    SoliditySmart ContractsdApps

How I handle projects

A sequence that survives contact with production.

Seven steps, each with a deliverable you can see — no phase where progress is invisible.

  1. 01

    Understand

    Clarify the business goal, requirements, users and technical constraints before any architecture is drawn.

  2. 02

    Architect

    Define system architecture, technology stack, APIs, database design, AI approach and cloud infrastructure.

  3. 03

    Build

    Develop clean, maintainable, scalable solutions using proven engineering practices and reusable patterns.

  4. 04

    Test & Secure

    Validate functionality, integrations, performance, security and production readiness before release.

  5. 05

    Deploy

    Implement CI/CD, cloud infrastructure, deployment automation and production configuration.

  6. 06

    Communicate

    Work to clear milestones with transparent timelines, regular updates and responsive communication.

  7. 07

    Optimize & Support

    Improve scalability, performance and reliability, and support post-launch requirements.

Selected work

Systems that went to production.

A patented product's revenue backbone, a ride-sharing platform's real-time layer, and enterprise features inside a Fortune 50 codebase.

QuickCam

Patented camera application · Backend architecture

StellarMind.ai · 2023–2025

Node.jsNestJSPostgreSQLAWSPayment gateways
Problem
A patented camera product needed a commercial backbone: paid tiers, reliable recurring payments, and a referral programme that could reward several levels of advocate without being gamed.
Approach
Architected and built the backend, owning subscription lifecycle management, automated payment processing and a multi-tier referral rewards system — plus the database design and third-party payment gateway integrations underneath them.
Outcome
A production revenue system handling the full subscribe → bill → reward cycle without manual intervention.

FlexxyDrive

Ride-sharing platform · Real-time systems

StellarMind.ai · 2023–2025

WebSocketsSocket.IONode.jsPostgreSQLAWS
Problem
A ride-sharing platform where drivers and riders had to stay in sync continuously — a domain in which delayed or dropped messages are product failures, not glitches.
Approach
Designed and built the real-time driver–user communication system on WebSockets and Socket.IO, engineered to hold many concurrent connections with low-latency delivery.
Outcome
Concurrent real-time communication running in production with low-latency message delivery between drivers and riders.

Enterprise platform at Johnson & Johnson

Fortune 50 enterprise application · Full-stack

Contract · Jun–Dec 2025 · Remote

ReactNode.jsCI/CDEnterprise securityAccessibility
Problem
Enterprise full-stack applications serving 22,000+ users, inside a large existing codebase governed by strict security, compliance and review standards where a careless change is expensive.
Approach
Developed and enhanced production features to established coding standards, branching strategy and CI/CD workflow, working with UX designers, backend engineers and product managers, with particular focus on accessibility, performance and usability.
Outcome
Shipped features to an application serving 22,000+ users within Fortune 50 governance, across the full lifecycle from requirements to production support.

Want the client side of the story?

Verified reviews, contract history and ratings are all on my Upwork profile.

See verified reviews

Experience

Seven years, four teams, one throughline.

From building client applications to leading a 15-engineer team — and a contract inside a Fortune 50 enterprise in between.

  1. Technical Lead

    StellarMind.ai

    Jan 2026 — Present

    Full-time · Ahmedabad, India

    • Own technical architecture and end-to-end delivery across multiple client engagements, translating business requirements into scalable system designs and production deployment strategies.
    • Lead and mentor a 15-member engineering team across frontend, backend and DevOps, driving technical execution, code quality and delivery standards.
    • Reduced average project delivery time by ~50% by standardising architecture templates, CI/CD pipelines, reusable component libraries and AI-assisted development workflows.
    • Run code reviews, technical design reviews and sprint planning, keeping projects aligned to quality and delivery goals.
    • Drive adoption of AI-assisted software development across the engineering workflow to accelerate development, debugging, testing and delivery.

  2. Full Stack AI Engineer

    Johnson & Johnson

    Jun 2025 — Dec 2025

    Contract · Remote

    • Contributed to enterprise-scale full-stack applications serving 22,000+ users, working within large production codebases and established engineering standards.
    • Developed and enhanced production features following strict coding standards, code review practices, branching strategies and CI/CD workflows.
    • Built and optimised frontend experiences with a strong focus on accessibility, performance and usability.
    • Worked within enterprise security, compliance, testing and deployment protocols in a Fortune 50 environment.
    • Collaborated with UX designers, backend engineers and product managers across the full development lifecycle.

  3. Full Stack AI Developer

    StellarMind.ai

    Mar 2023 — Jun 2025

    Full-time · Ahmedabad, India

    • Built and shipped 12+ client-facing web applications end to end — architecture, development, deployment and production monitoring.
    • Architected the backend for QuickCam, a patented camera application, including subscription lifecycle management, automated payment processing and a multi-tier referral rewards system.
    • Built the real-time driver–user communication system for FlexxyDrive using WebSockets and Socket.IO, supporting concurrent connections with low-latency delivery.
    • Designed and managed AWS infrastructure across EC2, S3, CloudFront, Lambda and RDS.
    • Integrated payment gateways, OAuth/social authentication and notification systems.

  4. Full Stack Engineer

    UEX

    Aug 2019 — Jan 2023

    Full-time · Ahmedabad, India

    • Designed, developed and delivered scalable web applications and SaaS platforms across frontend, backend, databases and cloud infrastructure.
    • Built responsive interfaces with React and Next.js, and scalable backend services and REST APIs with Node.js, NestJS and Python.
    • Implemented subscription billing, automated payment workflows and real-time communication features.
    • Deployed applications on AWS and maintained CI/CD pipelines.
    • Improved application performance, security, scalability and reliability.
EducationB.Sc. Computer Science, Ganpat University2016 — 2019

Technical expertise

The stack I build on.

AI / GenAI

  • AI Agents
  • RAG Systems
  • LLMs
  • OpenAI API
  • LangChain
  • AI Chatbots
  • Document Intelligence
  • Workflow Automation
  • Generative AI

Full-Stack

  • Python
  • TypeScript
  • React
  • Next.js
  • Node.js
  • NestJS
  • PostgreSQL
  • REST APIs
  • WebSockets
  • Socket.IO

Cloud & DevOps

  • AWS
  • EC2
  • S3
  • CloudFront
  • Lambda
  • RDS
  • Azure
  • GCP
  • CI/CD
  • Serverless

Blockchain / Web3

  • Solidity
  • Smart Contracts
  • dApps
  • Blockchain Integrations

Why clients work with me

The record, stated plainly.

  • 7+

    Years of software engineering

    Shipping production software since 2019, across startups, agencies and a Fortune 50 enterprise.

  • 15

    Engineers led

    Currently leading a 15-member engineering team across frontend, backend and DevOps.

  • 12+

    Applications delivered

    Client-facing applications taken from architecture through to production and support.

  • 22,000+

    Enterprise users supported

    Enterprise experience inside strict security, compliance and CI/CD standards.

  • ~50%

    Faster delivery

    Achieved by standardising architecture templates, CI/CD pipelines, reusable component libraries and AI-assisted workflows.

  • Fortune 50

    Enterprise environment

    Delivered inside Johnson & Johnson's security, compliance, code review and CI/CD governance.

  • Security, scale and production readiness

    The non-functional requirements are treated as requirements, not as an afterthought before launch.

  • Clear communication and end-to-end ownership

    Defined milestones, transparent timelines and responsive updates — with ownership that runs through production.

My working style

“Successful projects require more than good code.”

I focus on understanding the business objective, designing the right architecture, executing against clear milestones, communicating consistently, and taking ownership through production.

Questions

What clients usually ask first.

What kind of AI projects do you take on?

Production AI systems rather than experiments: AI agents, RAG systems over your own documents and data, LLM features inside existing applications, AI chatbots, and document-heavy workflow automation. The common thread is that the result has to run reliably against real business processes, APIs and SaaS platforms — not just demo well.

Can you take a prototype to production?

Yes — it is the most common reason people reach out. A working prototype and a production system differ in error handling, evaluation, cost control, latency, security and observability. I start by mapping what the system must guarantee, then design the architecture and build toward it, so you know what shipping actually requires before work begins.

How do you charge, and how do engagements start?

Engagements run through Upwork, where contracts, milestones and payments are handled and my review history is public. Most projects start with a short scoping conversation: you describe the workflow or problem, and I come back with an architecture, a milestone plan and a realistic timeline before any commitment.

How long does a typical AI project take?

A focused integration — adding a RAG-backed assistant to an existing product, say — is usually a few weeks. A full AI-powered SaaS build with subscriptions, payments and cloud infrastructure runs longer. I work in defined milestones so you see something working early rather than waiting for one large delivery at the end.

Do you work with existing codebases and teams?

Yes. I spent seven months on contract at Johnson & Johnson working inside enterprise applications serving 22,000+ users, under strict security, compliance, code review and CI/CD standards. I also lead a 15-member engineering team day to day, so joining an established codebase and its conventions is routine.

Which technologies do you build with?

For AI: OpenAI API, LangChain, RAG pipelines and agent frameworks. For applications: Python, TypeScript, React, Next.js, Node.js, NestJS and PostgreSQL. For infrastructure: AWS — EC2, S3, CloudFront, Lambda and RDS — with CI/CD and serverless architecture. I also build blockchain integrations with Solidity when a project calls for it.

Where are you based, and which time zones do you work across?

I'm based in Ahmedabad, India (IST) and work with clients worldwide, including US and European time zones. Remote collaboration is how I have worked throughout, including the Johnson & Johnson contract.

Can you review an existing AI implementation?

Yes. If something is built but unreliable, expensive or slow, an architecture review is often more valuable than a rewrite. That typically covers retrieval quality, prompt and model strategy, evaluation, cost per request, latency and failure modes, with concrete recommendations you can act on.

Get in touch

Have an AI idea, an existing product, or a technical challenge?

Send me your requirements and let's turn it into a scalable, production-ready solution. I typically reply the same day.

Hire me on Upwork WhatsApp

Opens WhatsApp or your mail client with the message pre-filled.

Writing

Notes on architecture and AI engineering.

All writing